Anw, having taken a look at the 2 lenses, they have different coatings, which is probably due to advances in coating(or cost cutting). the MIJ is ≈450g, while the MIT is ≈390g both without lens caps or filters.
Turning the focus ring form infinity to CFD makes the slotted screw drive turn ≈5.6 rounds on the MIJ, but only ≈4.6rounds on the MIT. this means(in theory) that the MIJ has a higher degree of accuracy while the MIT has a faster speed.. the MIT does indeed go from ∞-closest-∞ faster by about 1/5second. how much this affects the accuracy im not exactly sure and both lenses seem to acquire subjects just as fast.. have yet to look into accuracy or tracking. turning the focus ring in MF, the MIJ is quieter than the MIT, more rattling in the MIT even though its only purchased this year. however the sound difference is not noticeable in AF.
the back 'hood', which isolates the optics from the mechanical parts within the lens is longer in the MIT and at no point can you see the mechanics. the MIJ one is shorter and when the focus is at closest distance, the aperture mechanism can be seen inside. this in theory, means that the MIJ version is more susceptible to flare at close focus distances which would reduce contrast. i have not tried testing them in real world situations and this is just a possible effect. if it matters, the MIJ one is plastic while the MIT one is metal.
From what i've heard and somewhat believe, the MIJ lenses are made to tighter tolerances and have better optics(i have no way to actually test this, its just what i ve heard..) however, they lack the D chip which makes you have to choose D chip, or 'better build quality'. the D chip influences 3D matrix metering and flash metering(TTL-BL)..
what i would like to know is if the D chip also influences the accuracy of AF tacking... I have no evidence that it does, but no concrete quote that it does not.. my guess is that it does not so if you are careful with your metering and do not play with flash too much, the D chip makes little difference. I use manual lenses all the time and they dun allow 3D matrix metering anyway so little loss to me.
*this part is purely speculation*
it is possible that the 'big' difference in weight comes from the helicoid assembly used to move the rear focusing group. i know that this assembly is plastic in the 50 1.8D, and it is possible that the MIJ version uses a metal helicoid while the MIT uses a plastic helicoid. if thats the case then the MIJ versions are indeed built to a higher quality and will last longer than the MIT counterparts. but this shouldn't be too much of an issue and we don't see the MIT lenses just falling apart..
what we should be concerned with is the calibration of the lens groups within the lens which might be done to a higher tolerance int he MIJ that MIT which would thus lead to higher image quality. i ll be testing them side by side soon just for fun.